RAPID COMMUNICATION

In Situ Expression of a B7-Like Adhesion Molecule on Keratinocytes front Huntan Epiderntis

Thomas E. Fleming, William S. Mirando, Uwe Trefzer, Karen A. Tubesing, and Craig A. Elmets Department of Dermatology, Case Western Reserve University, Cleveland, Ohio, U.S.A.

The B7 adhesion molecule, a member of the immunoglobu­lin superfamily, has previously been identified primarily on cells of hematopoietic origin. Because B7 has been shown to facilitate interactions with T cells and because cells of the epidennis are proficient at binding and activating T lympho­cytes, studies were performed to determine whether B7 was expressed in human epidermis . A subpopulation of brightly staining B7-positive cells was observed in situ in normal human epidermis. Flow-cytometric examination of epider­mal cell suspensions that had been cultured for 24 h or longer

Inflammatory and immunologic reactions within the skin are highly regulated processes that are mediated to a large extent by chemotactic factors, cytokines, and cytokine receptors. Receptor-ligand interactions between adhesion molecules present 011 the surface of different cell types are also impor­

tant for the development, progression, and resolution of these reac­tions. Adhesion molecules mediate their effects by regulating cellu­lar migration patterns, promoting homotypic and heterotypic cell binding, and facilitating cellular communication among different cell types [1]. Keratinocytes have been observed to express at least two adhesion molecules that support T-cell interactions with the epidermis. These are intercellular adhesion molecule-1 (ICAM-1)

. and lymphocyte function-associated antigen-3 (LFA-3), which bind to LFA-1 and C02, respectively, on the surface ofT cells [2]. ICAM-1 is not expressed on normal keratinocytes, but can be ex­pressed in situ during inflammatory conditions and on cultured ker­atinocytes following treatment with tumor necrosis factor-alpha (TNF-a) and interferon-gamma (IFN-y) [3- 5]. Expression of this molecule has been shown to facilitate binding and signal transduc­tion between LFA-1-positive T cells and keratinocytes [5] . Cell­surface expression ofLFA-3 has also been detected on keratinocytes. Its role in cutaneous T-cell inflammatory processes is less well un­derstood. However, it does serve as an adhesion molecule to which C02-positive T lymphocytes can attach.

B7, a 44/54-kOa member of the immunoglobulin supergene family, has been recognized to be an adhesion molecule on B lym­phocytes [6,7] . It has also been referred to in the literature as BB-1 and B7/BB-1. Its natural ligands are C028 and CTLA-4 on T cells [8-10] . B7/BB-1 has been shown to stimulate Tcells by binding to the C028 molecule [11], and this interaction leads to increased T-cell proliferation and cytokine production [12,13] . Besides these quantitative effects, signal transduction through C028 stimulates T cells in a qualitatively unique way in that interleukin-2 secretion

Manuscript received Febrnary 15, 1993; accepted for publication June 20, 1993.

Reprint requests to: Dr. Craig A. Elmets, Department of Dermatology, Case Western Reserve University, 2074 Abington Road, Cleveland, OH 44106.

demonstrated that between 10 and 40% of cells expressed B 7 or a closely.related antigen. Il11muno:lec~ron microscopy , double-stall1ll1g procedures, and eXamll1atlOn of epidermal suspensions depleted of Langerhans cells all confirmed that the B7-positive cells were keratinocytes. These studies iden­tify human epidermal keratinocytes, a non-hematopoietic cell population, as a cell type capable of expressing a B7 -like adhesion molecule. Key words: CD28/Langerhans cells/ keratinocyte. ] Illvest Dennatol1 01:754 - 758, 1993

initiated via the C028-B7 interaction is not inhibited by cyclo_ sporin A [14]. There is evidence that the receptor-ligand interaction with B7 promotes B-cell differentiation as well [13]. B7 has also been detected on activated monocytes [8], on dendritic cells [15,16]. and on munne pentoneal exudate cells [17] . Recently. expression of a B7-like molecule has been demonstrated in lIitro on activated cultured keratinocytes [18,19] and on human Langerhans cells ~ epidermal cell cultures [20].

The focus of our experiments was to determine whether B7 wa present in situ in the epidermis and, if so, to determine the cell type expressing the B7 molecule. Because other investigators have found that B7 is involved in the T-cell activation process, and, within the epidermis, antigen presentation is largely attributed to Langerhans cells, it seemed likely that this cell type might express B7. We were surprised to find that B7 was present on epidermal keratinocytes and was not expressed by freshly prepared human Langerhans cells.

MATERIALS AND METHODS

Antibodies Purified BB-l, a murine monoclonal antibody of the IgM class specific for the B7 antigen, was the generous gift of Dr. Edward A. Clark, University of Washington, Seattle, W A [6]. The antibody was puri. fied from ascitic fluid by ammonium sulfate precipitation. In some experi_ ments, BB-l was conjugated with fluorescein isothiocyanate (F1TC) and then used for staining purposes. Other primary monoclonal antibodies used were anti-HLA-DR (Becton Dickinson Co., Mountain View, CAl. anti. COla (OKT6, Ortho Immune Diagnostic, Raritan. NJ), anti-desmogl~ (Boehringer-Mannheim, Indianapolis,IN),lgM (Calbiochem Corporation La Jolla, CAl, colloidal gold- and F1TC-col~ugated anti-mouse IgM (Ch'; micon, Inc., Temecula, CAl, and anti-mouse IgG-FITC (Chem.icon. Inc. Temecula, CAl.

E" face Staining of Epidermis Sheets .of epidermis were obtained by removlDg the roofs of vacuum-lIlduced suctIOn blIsters from the forearms of human volunteers as previously described [21]. The sheets were fixed in cold acetone, washed in phosphate-bnffere.d sa line (PBS), and then incubated with the appropriated primary monoclonal antibody (MoAb) for 18 h at 4°C. Primary antibodies used were BB-l (7.4 J.lg/ml), non-cross-reacti\'~ IgM (7.4 /lg/ml), anti-CDla (5 /lg/ml), and anti-HLA-DR (10 J.lg/rnI). The sheets were again washed in PBS to remove unbound antibody and wm­then treated with either FITC-conjugated anti-mouse IgM (100 J.lg/m1) at IgG (100 /lg/ml) , depending on the immunoglobulin class of the prima!}'

0022-202X/93/S06.00 Copyright © 1993 by The Society for Investigative Dermatology, Inc.

754

VOL. 101, NO.5 NOVEMBER 1993

a b c

Figure 1. Staining of epidermal sheets. Horizontal sections of normal human epidermis were incubated with (a) BB-l, 7.4 Jig/ml, (b) anti-COla (OKT6).? ,ug/ml, or (c) non-cross-reactive IgM, 7.4Jig/ml. The secondary anribody In a ll three tissue preparations was FITC-conJugated antl-mouse IgM. Bar, 25 Jim.

anribody. The sheets were again washed, mounted on sl ides with glycerine and examined with a Nikon immunofluorescence microscope.

Preparation of Epidermal Cell Suspensions Fresh human cadaver skin was obtained within 24 h of death and was taken only from individuals with no known infectious disease who were under 40 years of age. Subcuta­neous fat was removed from the skin specimens, which were then cut into 2 X 10 mm strips. The strips were then placed dermal side down in 0.5% dispase (Boehringer-Mannheim, Indianapolis, IN) and incubated for 24 h at 4' C. At the end of the incubation period, epidermis was separated from dennis and was placed in 0.15% trypsin (ICN Biochemicals, Inc., Cleveland, DH) for 60 min at 37'C. At the end of the trypsinization period, the disaggregation process was completed by vigorously pipetting the specimen for several minutes. 0.25% DNAse (Sigma Chemical Co., St. Louis, MO) was then added for 5 min, after which the cells were washed in culture medium. Cell viabi lity, as determined by trypan blue exclusion, was rypi­cally greater than 90%.

Staining ot Epidermal Cell Suspensions for Flow Cytometry Freshly prepared epidermal cel ls or cells that were incubated for various periods of time (as indicated in the Results) were stained with FITC-labcled 8B-l or non - cross-reactive FITC-Iabeled IgM. As a positive contro!' Ep­stein-Barr virus (EBV)-transformed B cel ls were stained with the same antibodies, Incubation was for 30 min at 4 'C using an antibody concentra­tion of 7.4 ,ug/m!' Fluorescence measurements were made on a flow cyto­meter (Cytofluorograph lIs, Ortho Instruments, Westwood, MA) using the 488-nm line of an argon laser operating at 200 to 300 mW. Propidium iod ide (1 J.lg/ml) was added to each sample 5 min prior to fluorescence-2Ctivated cell sorter to exclude dead cells.

Two-Color Analysis of Epidermal Cell Suspensions for Immunoflu­orescence Microscopy Epidermal cells were incubated with the antibody combinations 13B-l/anti - HLA-OR, BB- l/anti-COla, and BB-l/anti-des­moglein each for 30 min and 4 'c. Antibody concentrations used were BB-l, 7.4 J.Lgjml; anti-HLA-DR. 10 Jig/ml; anti-COla, .5 Jig/m.i; and anti-?es­rnoglcin, 10 ,ug/m!' The cells were washed twice 111 10% culture medIUm and incubated with a secondary anti-mouse IgG (100 Jig/ml) or IgM (100 Jig/m!) conjugated to either FITC or phycoerythrin (PE) for 30 min at 4 ' c. Cells were "Washed twice in 10% culture medium and examined on an immunofl uorescence microscope using the appropriate filters for FlTC and PE.

Depletion ot Langerhans Cells from Epidermal Cell Suspensions Freshly prepared epidermal cell suspensions (t07/ml) were incubated for 30 min with anti-COla monoclonal antibodies at 4 ' C. The cel l suspensions were washed and then incubated on a rocker panel with goat anti-mouse IgG-coated Dynabeads (Oynal, Inc., Great Neck, NY). The bead to target ,dl rario was 10 : 1. Following incubation at 4 'C, beads and bead-coated Langerhans cells were removed from the cell suspension using a Magnetic Particle Concentrator (Oynal, Inc., Great Neck, NY).

Immunoelect:ron Microscopy Epidermal cell suspensions that had been incubated for 24 h were incubated with BB-l or nOll- cross-reacting IgM at

KF.RATINOCYTE EXPRESSION OF A B7-LIKF. ADHESION MOLECULE 755

No Incubation

100 EC .. 19M EC + BB-l/B7

100 a b

Qi .Q

E ::l 10 Z 10

a; ()

0 O+--r~~.-~--~ 100 200400 600 8001000 100 200 400 600 8001000

Immunofluorescence

48 Hour Incubation EC .. 19M EC + BB-l/B7

100 100 C d

Qi .c E ::l 10 10 Z

a; ()

0 0 100 200400 600 8001000 100 200400 600 8001000

Immunofluorescence

100 Depleted EC .. 19M Depleted EC + BB-l/B7

100

Qi e .c E ::I 10 10 Z

a; ()

0 0 100 200400 600 8001000 100 200400 600 8001000

Immunofluorescence

Figure 2,. Flo,,:, cytometry. Cell suspensions were prepared from normal human epidermiS and stamed for flow cytometric analysis using either non­cross-reactive IgM (a , c, and e) or BB-l (b, d, andf) as the primary anribody. a,b) Cell suspenslOlIs stained immediately fo llowing disaggregation. c,d ) Cell suspensIOns mcubated for 48 It prior to staining. eJ) Cell suspensions depleted of Langerhans cells and incubated for 48 h prior to staining.

7.4 Jig/ml for 30 min at 4 'c. After washing twice, the cells were incubated with gold-labeled anti-mouse IgM for 30 min at 4' C and were aga in washed. The cell pellet was fixed with 3% glutaraldehyde in 0.1 M caco­dylate buffer. It was then embedded in a polymer resin, sectioned, and examllled on a Jeol 100CX transmission electron microscope.

RESULTS

I~ situ Ident!fication o~ B7-Positive Cells in en face Pre para­bO.ns of E~)1dermal Tlssue Sheets of normal epidermis were staIned for lmmUllOfluorescence microscopy with BB-1 to deter­mine whether this molecule was expressed ill situ. Substantial num­bers of B7-positive cells were identified (Fig 1a). The positively stained ~ells w~re present in clusters and had a rounded morphol­ogy. TIllS stallung 'pattefl~ contrasted with that seen when epider­mal sheets were stamed With antl-CD1a MoAb to detect epidermal Langerhans cells (Fig 1b). With anti-CD 1 a, positively stained cells were diffusely distributed and exhibited a dendritic morphology charactenstlc of Langerhans cells. Because £B-1 is an IgM class antibody, control epiderma l sheets were stained using non-cross­reacting IgM as the primary reagent. Only background staining was observed (Fig le).

756 FLEMING ET AL

Table I. Flow-Cytometric Analysis of Epidermal Cells Stained for B7'

Cell Type

Epidermal cells, 48-h incubation EBV-transformed B cells

Mean Fluorescence Intensity

Anti-137 (BB-l)

49.6 38 .9

IgM Control

3.1 8.2

• Cells were stained with either BB-! (7.4 /lg/ml) or murine non -cross-reactive IgM (7.4 /lg/ml) as the primary antibody. The secondary antibody in both cases was FITC­conjugated anti-mouse IgM (100 pg/m l). Mean fluorescence intensiry is expressed in arbitrary units.

Identification of B7-Positive Cells in Suspension of Epider­mal Cells B7-positive cells could also be identified in epidermal cell suspensions following trypsin disaggregation. In most in­stances, few, if any, B7-positive cells could be detected immediately after trypsinization (Fig 2a,b). Expression of the B7 antigen became apparent after 24 h and was upregulated when cells were cultured for longer periods of time. After 48 h in culture, a large proportion of the cells, ranging from 10 to 40%, was B7 positive (Fig 2c,d) . EBV -transformed B cells were used as a positive control (Table 1).

The large percentage of epidermal cells staining positively for B7 suggested that keratinocytes were the predominant cel! type to ex­press this molecule. To examine this issue further, HLA-DR­positive Langerhans cells were removed from cell suspensions prior to the staining procedure. Preliminary studies established that the depletion procedures rendered cell sllSpensions devoid of CDla­and HLA-DR-positive cells (data not shown). No difference in fluorescence intensity among depleted and non-depleted epidermal cell popnlations was observed (Fig 2e,J). Similar results were ob-

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

tained when anti-CDla MoAb was employed to remove Langer­hans cells from the preparations (data not shown) .

Immunoelectron Microscopy lml11unoelectron microscopy was performed to determine definitively the type of cell expressing the B7 adhesion molecule. Unfractionated e~iderl11a l cell suspen­sions that had been cultured for 24 h were first Illcubated with BB-l and were tben treated with go ld-labeled anti-mouse IgM. Gold beads cou ld be easily identified on the surface of ce lls (Fig 3, inset) . All ceUs labeled with BB-l possessed abundant keratin filaments and desl11osomes (Fig 3). None of the cells examined had the ultra­structural characteristics of Langer hans cells or melanocytes. More­over, gold-labeled cells were not observed when a non- cross­reacting IgM antibody was substituted as the primary reagent (data not shown).

Double Staining of Epidermal Cell Suspensions Keratino_ cytes cultured for 48 h were stained with both BB-l and anti-des_ mog~ein, a ~esmo~omal protein. present only on keratinocytes wlthm the ep~dermls (~lg 4). Prelllmnary .tests established that IgG secondary antibodies did not cross react With IgM primary antibod_ ies and that IgM secondary antibodies did not cross react with IgG primary antibodies (data not shown). Approximately half of the desmoglein-positive cells were B7 positive, providing further evi_ dence that keratinocytes express B7. HLA-DR-positive and CD la-positive Langerhans cells did not stain positively for BB-! in double-staining studies (data not shown).

DISCUSSION

These studies demonstrate that a B7-like adhesion molecule can be found i,l situ 011 epidermal cells, specifically human keratinocytes, This was definitively demonstrated by two-color immunofluores_

Figure~. lmmunoelectron microscopy. Epidermal cell suspensions were incubated with BB-l followed by colloidal gold-conjugated anti-mouse IgM. A representative cell is shown here (magnification X 5000). Labeled antibody (",row) bound to the cell surface can be seen on the il/se! (il/se! enlarged 4X). Bar, _ j.Lm.

VO L. 101. NO. 5 NOVEMBER 1993 KERATINOCYTE EXPRESSION OF A 137-LIKE ADHES ION MOLECULE 757

a b c

Figure 4_ Two-color immunofluorescence microscopy using BB-l and anti-desmogJein. Epidermal cell suspensions were stained for B7/ BB-l (FITe-conjugated secondary antibody) and antl-desmogletn (PE-conJu-ated secondary antibody). a, b, and c are from the same field and tillS field IS

~epresen.t:~t:ive of the cells exa~ined. a, B7/BB-l- positive cells. b; Desmog­lein-posltlve cells. c, normal hght. Bar, 25 J.1.m .

cence when B7-positive cells were observed to be desmoglein-~osi­tive and by immunofluorescence microscopy when B7-posltlve cells were found to contain the ultrastructural components of kera­tinocytes . More importantly, this B7-l ike adhesion molecule was expressed in situ in epidermal sheets , suggesting that the molecule may h ave an important ill llillO role in keratinocyte - T cell interac-

tions. T he function of this B7 -like mol ecule on keratinocytes is open to

speculation. For B cells [12,13] and murine dendritic cells [15], B7 is an important costimulatory molecule for T-cell activation. B7 may play a similar function on keratinocytes. Although Langerhans cells, rather than keratinocytes, have traditionally been considered to be the major antigen-presenting cell of the epiderml.s, keratll1o­cytes can, in certain circumstances, perform this funct10n as ,:,ell. For example, keratinocytes induced to express class II major hlst~­compatibility complex molecules by IFN-y can stimulate allogeneic T cells to proliferate when interleukin-2 is added to the cultures [2223]. Phorbol myristate ace tate- treated keratinocytes have al~o bee~ shown to be effective antigen-presenting cells for allogeneic responses without the addition of exogeI~ous cytokmes [24]. Pre­liminary studies from our laboratory II1dlcate that epidermal cell lines devoid of Langerhans cells provide all the necessary accessory signals for T-cell activation by the bacterial superantigen staphylo­coccal enterotoxin B. The addition of anti-B7 antibodies substan­tially inhibits T-cell proliferation in this system (T. E. Flemll1g and C. A. Elmets, manuscript in preparation).

B7 may be important for other bio logic activities within the epidermis .as well. Specifically, B~ may serve as a homing receptor for migrating T lymphocytes and Its expressIOn may be enhanced 111

cutaneous diseases associated with an inflammatory infiltrate. B7 expression may be particularly relevant for recruitment of cytotoxic T cells to the epidermis because B7 has recently been found to be a ligand for CTLA-4, a molecule found on a subpopulation of cyto­toxic T cells [9] .

Our inability to identify B7 on epidermal Langerhans cells was based on our observations that no cells were both HLA-DR positive 2nd B7 positive, that all B7-po~itive cells stained with ?esmoglein anribodies, and that all B7-posltlve cells contamed kcratm filaments and desmosomes by immunoelectron microscopy. This findin g was somewh at surprising because cultured murine Langerhans cells have been shown to express this molecule [15]. It is important to mess that our studies do not exclude the possibility that human Langerhans cells express the B7 molecule under special conditions. Langerhans cells have been show~ to ~lcrease th.ei~ .cell-surface expression of CD54 and class II major 111StoCOmpaubility complex

determinants w hen placed in culture [25 ,26]. In murine systems such changes have been associated with an enhanced capacity to present antigen to T lymphocytes [27]. It is certainly possible th at cultured human Langerhans cells also begin to express B7 under these conditions. A second possibili ty is that the BB-l monoclonal antibody recognizes an epitope on a B7-like molecule expressed by keratinocytes and that this epitope does not exist on Langerhans cells expressing a different B7-like molecule. This second possibil­ity is supported by two recent reports. In one study, activated, cu l­tured keratinocytes stained positively for BB-l but not with other anti-B7 monoclonal antibodies [1 9]. In the second study, investiga­tors using a CTLA-4-Ig fusion protein have detected B7 on cul­tured Langerhans cells but not on cultured keratinocytes [20] .

The role of th e B7 adhesion molecule in cutaneous immunobiol­ogy and in the pathology of cutaneous inflammatory disease re­n~ains t~ b.e determined. A number of skin diseases including aller­gic and Irritant contact dermatitis, psoriasis, bullous drug eruptions, and graft-versus-host disease are associated with a lymphoid infil­trate in the epidermis. Although other adhesion molecules have been shown to facilitate T -cell binding to the epidermis, the consti­tutive expression of a B7-like molecule on keratinocytes suggests that it may play a ro le as well.

In summary, keratinocytes express an adhesion molecule ill sitll that shares binding domains with B7 based on cross reactivity with the anti-B7 monoclonal antibody BB-l. This mol ecule is either B7 itself or a closely related molecul e. B7 has been identified on various cell types, primarily of hematopoietic origin, and costimulates T cells via interaction with CD28 and CTLA-4. Its role in cutaneous immunobiology and in the pathology of cutaneous inflammatory and immunologic diseases remains to be determined.

WeackrlOwledge the /eelll/ ica l expertiseofHaipitlg Tallg mtd the secreta rial assistall ce of Carol Highslllith . We wish to /llOIIk David R. Kaplall , M.D., Pt..D. olld W. Hmry BOOIII, M.D.for theircareflll review of the lIIatlllScript, alld weare illdeb/ed /0 Edward A. Clark, Plt.D. for provision of BB-1 1II0lloclollai all/ ibody. We also tltallk Dr. Carlos Stl ll/OSCOY for Itelp ill prOCllrelllell/ of tiss lle.

Tltis 1II0rk was sllpported by NIH gra II ts A R32593, AR01765, CA43703, alld AR39750, IIIld by flillds frOIll the Proctora"d Cam bleCompallY. M.r. Flemillg is the recipietJt of all A merican Derlllatological Associatioll lIIedica l sllldelli jellowsitip. Dr. Trcfze r is slIpported by 0 f ellowship g ratH frO Ill the Dermatology FOlll/datioll .

REFERENCES

1. Springer TA: Adhesion receptors of the immune system. Nawrc 346:425-434 , 1990

2. Singer KH , Le PT, Denning SM. Whichard LP, Haynes BF: The roleof adhesion molecules in cpithclial - T cell interactions in th ymus and skin.] Invest Derma/ol 94:85S-90S, 1990

3. Singer KH, Tuck DT, Sampson HA, Hall RP: Epidenml keratinocytes express the adhesion molecule intercellular adhesion molcculc-l in inflammatory der­matoses.) IrIVes( Der",a(0192:746 -750, 1989

4. NickolofI Bj , Griffiths CEM, Baadsgaa rd O. Voorh ees Jj, Hanson CA, Cooper KD: Marked ly diminished epidermal ker3tinocyte expression of intercellular adhesion molecule-! (ICAM-1) in Sczary syndrome. )AMA 261:22 17-222 1. 1989

5. Dustin ML, Singer KH, Tuck DT, Springer TA: Adhesion ofT Iymphoblasts to epidermal keratinocytes is regulated by interferon gamma and is mediated by intercellular adhesion molecule-I (I CAM-l).J ExpMed 167:1323- 1340, 1988

6. Yokochi T, Holly RD, Clark EA: 13 lymphoblast antigen (13B-1) expressed on Epstein-Barr virus-activated 13 cell blasts, B Iymphoblastoid cell lines, and Burkitt's lymphomas. ) ]"''''''''011 28:823-827. 1982

7. Freeman G, Freedman AS, Segi lJM, Lee G, WhitmanJF, Nadler LM: 137, a new member of the Ig superfamily with unique expression on activated and neoplas­tic 13 cells. ) /"''''''''0/143:2714-2722, 1989

8. Freedman AS, Freeman G, Rhynharr K, Nad ler LM: Selective induction of 137/ BB-l on inrcrfcron-i' stimulated l11onocytes: a porentialmcchanism for ampli­fication of T cell activation through the CD28 pathway. Cell ]",,,,,/1101 137:429-437,1991

9. Linsley PS, GreeneJL, Tan P, Bradshaw J, LedbetrcrJA , Anasetti C, Dank NK: Co-expression ano functiona l cooperation of CTLA-4 and CD28 on actinted T lymphocytes.) Exp M cd 176: 1595 - 1604, 1992

10. Valle A, Aubry JP, Durand I, Banchereau J: IL-4 and IL-2 upregulate the expres­sion of antigen B7, the B cell counterstructure to T cell CD28: and amplifica­tion mechanism for T-.B cell interactions. 1m ImmrmoJ 3:229-235, 1991

11 . Linsley PS, Clark EA. Ledberter JA: T ce ll-antigen CD28 med iates adhesion with

758 FLEMING ET AL

B cells by interacting with activation antigen B7 /BB-I. Proc Nat! Acad Sci USA 87:5031- 5035, 1990

12. Koulova L, Clark EA, Shu G, Dupollt B: The CD28 ligand B7/BB-l provides costimulatoey signal for alloactivation of CD4+ T cells. ] Exp Med 173:759-762, 1991

13. Linsley PS, Brady W , Grosmaire L, Arulfo A, Damle NK, Ledbetter JA: Binding of the B cell activation antigen B7 to CD28 costimulates T cell proliferation and intcrleukin 2 mRNA accumulation.] ExpMed 173:721-730,1991

14. Bloemena E, Van-Oers RH , Weinreich S, Srilma-Meinesz AP, Schellekens PT, Van-Lier PT: The influence of cyclosporin A on the alternative pathways of human Tcell activation ill vitro. Ellr] Immll,10119:943 -946, 1989

15. Larsen CP, Ritchie SC, Pearson TC, Linsley PS, Lowey RP: Functional expression of the costimulatoey molecule, B7/ BB-l , on murine dendritic cell populations. ] ExpMed 176:1215 - 1220, 1992 .

16. Young JW, Koulova L, Soergcl SA, Clark EA, Steinman RM, Dupont B: The B7/BB-1 antigen provides one of several costimulatoey signals for the activa­tion of CD4+ T lymphocytes by human blood dendritic cells in vitro.] e lill IIIVest90:229-237, 1992

17. Razi-WolfZ, Freeman GJ, Galvin F, BenacerrafB, Nadler L, Reiser H: Expres­sion and function of the murine B7 antigen, the major costimularoey molecule expressed by peritoneal exudate cells. Proc Nail Acad Sci USA 89:4210 -4214 , 1992

18. Augustin M, Dietrich A, Niedner R, Kapp A, SchopfE, Ledbetter JA, Brady W, Linsley PS, SimonJC: Phorbol-12-myristate-13-acetate-treated human kerati­nocytes express B7-like molecules that serve a costimulatoey role in T-cell activation.] IrIVest Derlllatoll00:275 - 281, 1993

19. Symington FW, Brady W , Linsley PS: Expression and function ofB7 on human epidermal Langerhans cells.] IIIIlIlImoI150:1286-1295, 1993

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

20. Nickoloff BJ, Mitra RS, Lee K, Turka LA, Green J, Thompson C, Shimizu Y: Discordant expression of CD28 ligands, BB-l , and B7 on keratinocytes ill vitro and psoriatic cells illllilio. A lii] Pat!.01142:1029-1040, 1993

21. Sontheimer RD: The mixed epidermal cell-lymphocyte reaction. I. Human epi­dermal cells elicit a greater allogeneic lymphocyte response than do auto logous peripheral blood lymphoid cells.] IrIllIlIl,101130:261 2-2614, 1983

22. Morhenn VB, Nickoloff BJ: Interleukin-2 stimulates resting human T lympho­cytes' response to allogeneic, gamma interferon treated keratinocytes.] IrlVesr Dem,atol 89:464 - 468, 1987

23. Niederwieser 0, AubockJ, Troppmair J, Herold M , Schuler G, BoeckJ, Lotz J , Fntsch P, Huber C: IFN-med.ated mductlon of MHC antigen expression on human keratinocytes and its influence on in vitro aUoimmune responses. ] 111111111110/140:2556 -2564,1988

24. SimonJC, Cruz PO, Bergstresser PR, Davis LS, Tigelaar RE: Phorbol myristate acetate-activated keratinocytes stimulate proliferation of resting peripheral blood mononuclear lymphocytes via a MHC-independent, but protein kinase C- and intercellular adhesion molecule- I-dependent mechanism. ] 1m III '11101 146:476-484, 199 1

25. Simon jC, Cruz PO, Tigclaar RE, Sontheimer RD, Bergstresser PR: Adhesion molecules CO Il a, CD I8, and ICAM-l on human epidermal Langerhans cells serve a functional rol e in the activation of alloreactive T cells.] bIVest Dermatal 96:148-151, 1991

26. Romani N , Lenz A, Glassel H, Stossel H, Stanzl U, Majdic 0, Fritsch P, Schuler G: Cultured huma n Langerhans cells resemble lymphoid dendritic cells in phenotype and function.] b IVest D"III.toI 93:600-609, 1989

27. Shimada S, Caughman SW, Sharrow SO, Stephany 0 , Katz SI: Enhanced antigen presenting capacity of cultured Langerhans cells is associated wth markedly increased expression of la anti gen.] [111111'/110/139:2551 - 2555 , 1987